ENERGY Pipeline by Felipe Germini Podcast

About ENERGY Pipeline by Felipe Germini Podcast

Insider analysis on energy, markets, and geopolitics—and the power dynamics that quietly move prices. fgermini.substack.com (https://fgermini.substack.com?utm_medium=podcast)

Publisher

Felipe Germini

Category

news · news

Language

en

Explicit

No

First episode

26 May 2026

Latest episode

27 Jul 2026

Latest episodes

16 episodes in the feed.

1. 

   27 Jul 2026

   The Signal: The Barrel got cheaper but the Voyage did not.

   The board went red on Monday, and the desks read it as relief. Brent down 7.56 percent to $89.46. WTI at $83.43. Murban, the light Abu Dhabi grade that loads at Fujairah and prices off the Gulf, down 12.57 percent, almost double the Brent move. The tape said the war was ending. Then look one line down. Heating oil, the cleanest proxy for diesel on the screen, fell 1.35 percent. Crude lost more than seven percent and the distillate barely flinched. That gap is not noise. It is the whole story, and most of the tape missed it. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

2. 

   21 Jul 2026

   JPMorgan: The Bank Sold the Barrels. Now It Is Buying the Ships.

   ON JULY 16, Hanwha Ocean filed a disclosure in Seoul: two very large crude carriers, 394.3 billion won, an undisclosed North American buyer. The anonymity did not survive the week. Shipbroking desks named the buyer as JPMorgan, or more precisely the shipowning platform inside its asset management arm, and the two VLCCs at roughly $132.5 million each joined a slate that now runs to about ten supertankers across Chinese and South Korean yards, close to $1.3 billion of crude-carrying steel, deliveries stretching from 2029 to March 2030. Read that again slowly. The bank that sold its physical commodities business in 2014, under regulatory pressure and with visible relief, is now one of the largest single buyers of crude tankers in the sharpest ordering boom the market has ever recorded. Not oil. Ships. The distinction is the whole story. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

3. 

   16 Jul 2026

   Brazil Rare Earths: The 21 Million Tonne Illusion

   This week, the Brazilian Senate postponed the vote on the first law meant to do something with Rare Earths. In the meantime, a Texas company agreed to pay $2.83 billion for Brazil’s only producing rare earth mine. Reserves are geology. Supply chains are chemistry, capital and law. Only one of those three is settled in Brazil’s favor. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

4. 

   15 Jul 2026

   The Price Cap Just Gave Moscow a Raise

   The West is now running two sanction architectures on the same barrel: a European price cap that loosens itself when oil rallies, and an American tariff that pushes buyers away from the only system the cap can see. They are working against each other, and someone is going to trade the seam. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

5. 

   14 Jul 2026

   The Last Dance: Hormuz’s New Normal

   Late Monday, the President of the United States posted that the Hormuz Strait “is OPEN, and will remain OPEN, with or without Iran,” that America is reinstating “THE IRANIAN BLOCKADE,” and that the United States will henceforth be “THE GUARDIAN OF THE HORMUZ STRAIT,” reimbursed at 20 percent of all cargo value for the service. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

6. 

   13 Jul 2026

   America's 319 Million Barrel Problem

   The number printed this morning: 319.5 million barrels. That is what remains in the United States Strategic Petroleum Reserve as of the week ending July 3, and you have to go back to April 1983 to find less oil in those salt caverns. Ronald Reagan was in his first term. The reserve was still being filled for the first time. Against an authorized capacity of 714 million barrels, America's oil insurance policy is now 44 percent funded, and the tape it must be rebuilt against closed near $79 Brent today, up roughly ten percent in a week. The world's largest strategic reserve has become the world's largest short position in physical crude, and the delivery notices start arriving on November 1. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

7. 

   3 Jul 2026

   The Barrel Forgot the War. Brazil's Onshore Ledger Did Not.

   Front-month WTI, the August contract, settled under $68 this week. Dated Brent sits under $71, the weakest print since February 27, with the August CFDs telling the same flat story. February 28 was the day US and Israeli strikes shut the Strait of Hormuz and Brent went looking for $120. Four months later the tape has erased the entire episode, as if a war in the world’s most important chokepoint were a data glitch someone corrected. The war premium was rented, not owned. Markets do not price history. They price flow, and the flow never really stopped. What interests me is not the round trip itself. It is where a $68 barrel lands once it crosses the Brazilian ledger, because it lands twice, on opposite sides of the page, and only one of those entries is getting any attention. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

8. 

   1 Jul 2026

   Oilfield Services a Workforce between the Iron and the Algorithm

   Look at where the service majors are putting their money, not where they are pointing their press releases. SLB now runs a standalone Digital Division. Annual recurring revenue near 926 million dollars, compounding double digits quarter on quarter, the fastest-growing business in the house. Put it in proportion first, because the number alone will mislead you. That is a small slice of a company turning over tens of billions, two to three percent, not the whole income statement. The signal is not the size of the line. It is the direction the capital and the org chart are pointing, while the same company’s drilling customers idle rigs and trim exploration budgets with WTI stuck in the low sixties. Halliburton sells LOGIX automation and remote operations. Baker Hughes is chasing the same turn. All three are leaning into AI data-center infrastructure work because the old growth engine sputtered and the new one runs on software margins, not on bodies. Watch which team grows and which seat is never refilled after the next cut, and you have already read where the value is going. The digital desk hires. The field engineer’s chair sits empty. Nobody vanished in a press release. They became a cost line to be managed down. That is the validation. Now the part nobody on the technical career track wants to hear. 01 · THE SURVIVOR POPULATION AI is landing on a workforce two crashes already gutted This is the detail every comparison to banking or law gets wrong. AI is not arriving to a fat workforce in upstream. It is arriving to a survivor population, a crew already cut twice to the bone. Two price crashes did the heavy work before a single model touched a drilling program. When Brent broke in 2014, and again when demand fell off the table in 2020, the service majors shed people like ballast over the side of a sinking hull. SLB alone cut roughly 34,000 jobs across 2015 and 2016, about a quarter of its entire workforce, and discovered it could run the same scope of work with far fewer hands. The crews that came back came back thinner. They never came back to the prior peak. The schools tell the same story from the other end. US petroleum-engineering enrollment is down about 75 percent from its 2014 peak of 11,474 students. Texas Tech off 88 percent. Oklahoma off 90. A whole cohort looked at the price chart and the transition narrative and walked into a different building on campus. So the displacement arithmetic in this industry runs backward from the arithmetic everywhere else. The marginal person AI removes from a bank is one of thousands doing cognitive routine. The marginal person it removes from a service company is rarer, more expensive to replace, and carries knowledge that took fifteen years and three failed jobs to build. The fat is long gone. What is left is muscle and a thinning seam of grey hair that remembers, in its hands, why the last cement job channeled. Hold that picture, because it inverts the comfortable story. In most industries automation eats the factory floor first and the credentialed desk is the safe harbor. Upstream flips it. The hands on the iron are the hardest thing on earth to automate, fenced behind a safety culture where the bar to remove a human from a high-consequence task sits near impossible. The exposed roles are the desk-bound technical ones that always assumed a diploma was a moat. The model reads the log. The hand on the rig floor reads the well. Those are not the same skill, and only one of them is being commoditized this decade. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

9. 

   26 Jun 2026

   The Brake Just Came Off Subsea Pricing in Brazil

   On 23 June, Brazil’s CADE (antitrust body) cleared the combination of Saipem and Subsea7 with nothing attached. No divestiture. No behavioural undertaking. No asset carve-out. For anyone who prices the cost of putting steel on the seabed, that absence is the entire event. Petrobras, ExxonMobil and TotalEnergies spent nine months fighting this deal through Brazil’s industry body precisely to avoid that line. They wanted a structural brake on subsea installation pricing. They did not get one. The combination is real and it is nearly finished. Binding merger agreement signed July 2025. Both extraordinary shareholder meetings approved it on 25 September. Italy waved it through under golden-power review with strategic-asset conditions. The UK cleared it. Brazil was the last hard antitrust gate, and it fell on 23 June. What is left is listing mechanics on Milan and Oslo, with management still guiding to a second-half 2026 close. This is not a proposal to handicap. It is a done structure to position around. So stop modelling the holding company. The combined group will report roughly €21bn of revenue, more than €2bn of EBITDA, €43bn of backlog, 60-plus construction vessels and about 44,000 people. Impressive, and mostly noise for what matters here. Around 84% of combined EBITDA sits inside a single entity, the renamed Subsea7, a Saipem7 Company. That ring-fenced, UK-incorporated subsea business is the asset. The onshore EPC arm, the drilling rigs and the infrastructure scope are the wrapper. If your model runs at the group line, you are averaging the engine with the chassis. The gates are behind it The deal walked a long regulatory corridor and is now nearly through. The memorandum of understanding came in February 2025. The binding agreement followed in July. Shareholders approved in September, with Saipem clearing the whitewash majority it needed. Italy’s Council of Ministers cleared it under golden power. The UK competition authority cleared it. Then Brazil, the one everyone watched, cleared it unconditionally. Completion triggers change-of-control on the €500m 2029 convertible, which is a discrete refinancing item to model, not a deal risk. The Brazil question, sized Here is where I will not hand-wave, because the whole pricing argument turns on it. I rebuilt the Petrobras subsea market from the bottom up, contract by contract, off the public register, spreading each award across its term by days and attributing it to calendar years. It is one operator’s reconstruction, not an audited market share, and I will say that plainly. But it is built on real contracts, not on a consultant’s pie chart. What it shows: across the eight principal contractors that actually win Petrobras subsea work, this captured legacy pool peaked near $1.9bn a year in 2023. The two merging parties together averaged about 44% of that legacy eight-player pool across 2022 to 2024. Pre-merger they were the top two of eight, Subsea7 around 23% and Saipem around 21%, two separate disciplines on the tension line. Post-merger they are one block at roughly 44%, twice the size of the next rival. That is the number CADE could not pretend it did not see. One honest caveat, because it cuts against my own headline. That 44% is share of the legacy eight. It does not include Allseas, the entrant now winning awards, which means the true forward share is already below 44% and falling as the new scope lands outside the merged book. The market is concentrating and de-concentrating at the same time. Hold both facts: the merger creates the largest block this market has ever seen, and the data already shows that block past its peak share. Run the concentration test the regulators themselves use and it is not close. On this eight-player pool the Herfindahl index moves from roughly 1,600 before the merger to about 2,550 after. That is a jump of around 950 points, lifting the market from moderately concentrated to highly concentrated and crossing the 2,500 line that normally triggers a hard look. The change alone is several times the threshold that flags a deal for scrutiny. On vessels the picture is starker still. Eight of the twelve units worldwide capable of the most complex deepwater SURF work, the heavy J-lay and high-tension reel-lay hulls, would sit inside one firm, around two-thirds of the global high-spec fleet. Read that as global capability, not Brazilian inventory. Those hulls work the North Sea, West Africa and Brazil in the same year, and a vessel qualified for Petrobras is not the same as a vessel in Brazilian water next quarter. The scarcity is global, which is precisely why it travels. This is the stake. Petrobras is roughly 90% of Brazilian subsea demand, and Brazilian pre-salt is one of the densest deepwater SURF markets on earth. A buyer that large facing a vendor that concentrated, on a multi-year multi-billion pipeline, is the exact setup where a few points of installation price uplift turns into real money. On a $1.9bn annual market, the arithmetic is not subtle. The question for your desk is no longer whether the deal closes. It is where the pricing power sits the morning after, and what disciplines it now that the regulator chose not to. The rest is for paying subscribers This serves the equity and credit desks covering offshore services and the strategy teams inside the operators who have to budget pre-salt tie-ins for the next five years. Above the wall is the concentration. Below it is what it does. You get where the dominance actually is by segment, the three scenarios for SURF pricing with the level that confirms each, the read across to Guyana and to European offshore wind, and the watchlist with the specific tenders that will settle the argument. Comes with a board-format desk pack. Two exhibits to open first: the segment dominance grid that shows where Saipem7 is unbeatable and where it is absent, and the upcoming Petrobras SURF award calendar with the entrant already eating share. Subscribe to read the scenarios and the watchlist. Signal monthly $10. Flow + Horizon annual $100. The Desk tier at $300/year buys the models, the dashboards, the locked archive, and priority on reader questions, and it expenses as research. ENERGY Pipeline by Felipe Germini is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. Where the dominance actually is Now the part the 44% headline hides, and the part that keeps this from being a clean monopoly story. Concentration is not uniform across the work. Break the same contract data into the five things Petrobras actually buys and the picture splits hard. In heavy SURF and EPCI installation, the merged firm is roughly 79% of the captured market. That is the chokehold. In vessel charter and pipelay support it is around 69%, strong but contested. Then it falls away. In inspection, repair, maintenance, survey and diving, the merged firm is a minor player, and DOF and OceanPact lead. In subsea production systems, the trees and the hardware, Saipem7 is effectively absent, and that work belongs to TechnipFMC and OneSubsea. So the dominance is real but it is a spike, not a slab. It sits on the single discipline where high-spec vessels are scarce and substitution is hardest. The counter-pressure is already in the water If the bull case for pricing power were airtight, you would not see a third installer winning awards before the merger even closes. You do. Allseas, running its own vessel Audacia, has already taken Buzios 10 and Atapu 2. That is a credible entrant eating SURF share inside Petrobras’ own program while the antitrust ink is still drying. The honest qualifier, because one hull is not a market: entry here is slow, not easy. A complex SURF vessel is a multi-year build, and Petrobras qualification is its own gate. Allseas got through the door, but the next one cannot simply show up, which means the merged firm can still hold the whip on the next two or three super-major awards even with the door technically open. Add to that what Petrobras itself is reportedly studying: chartering large pipelay vessels directly rather than handing out full single-supplier EPCI packages. Unbundle the scope and you dilute the integrator’s grip by design. And then the structural counterweight nobody on the sell-side likes to model because it cuts the wrong way. Petrobras is not a fragmented set of buyers. It is close to a single buyer for the whole market. Monopsony sits on the other side of monopoly, and a buyer that controls 90% of demand has its own discipline to apply. That is most of why CADE felt able to clear without remedies. The combined firm is large, but it sells to one counterparty that can move a tender calendar, sponsor an entrant, and write 40-to-50% local-content rules into the bid. The pricing fight is real. It is not one-sided. What the synergies are, and when The company guides to around €300m of run-rate synergies, and the honest detail is in the timing and the mix. The largest bucket is fleet optimisation, better utilisation and positioning of a combined vessel book. Procurement is next, then sales and tendering rationalisation, then process. The full run-rate only lands in the third year after completion. This is not day-one accretive, and the biggest single lever is precisely the vessel utilisation that the entrant and the unbundling threaten. Capital return is set at a minimum 40% of free cash flow after lease repayment, with an investment-grade rating as a target, not a fact. The IG re-rating is the genuine equity story here, and it is a 2027-2028 question. Three scenarios for SURF pricing Uplift, the bull case for the contractor. Clearance without remedies holds, the entrant stays a fringe player, and on the largest ultra-deepwater packages the merged firm captures a few points of installation price uplift that compound across a multi-year book. The confirming signal: the next super-major Buzios or Sepia SURF award lands above the recent ~$1.25bn-plus run rate on comparable scope, and award share keeps concentrating. Watch the price per comparable scope, dated to the award month, not the headline contract value. Contested, my base case. Allseas and any second entrant stay live, Petrobras unbundles a meaningful share of pipelay from EPCI, and the buyer’s scale offsets the seller’s concentration. Pricing drifts up on cost-push, the 40-to-50% local content and deeper water, not on market power. The confirming signal: at least one of the open tenders, Sepia 2, the SEAP packages, or the 518 km flexible-pipe tender opened in January, goes to a non-merged party or to a direct-charter model. Reopened, the tail risk to the deal’s comfort. The clearance came from CADE’s General Superintendence, which means it sits inside the window where a third party can appeal or the Tribunal can call the case up for its own review. The movers here are not abstractions. They are the IBP and the same operators, Petrobras, ExxonMobil and TotalEnergies, that filed in the first place. The same buyer that fought the deal also sets every future tender. If award concentration spikes immediately and visibly, the pressure to revisit, or to extract behavioural undertakings after the fact, climbs with it. Low probability, real tail. The confirming signal is a Tribunal call-up or an operator filing, not a price print. The global ripple Brazil is the test case, not the boundary. The same eight-of-twelve high-spec vessel logic travels straight to Guyana, where ExxonMobil is running the other dense deepwater build-out of the decade, and where the same scarce units would have to show up. A pricing read that holds on Buzios informs every Stabroek tie-in budget. That is why Exxon filed in Brazil over a Brazilian deal. It was pricing its own forward book somewhere else. Second, the capital story is European. A combined group incorporated in Italy, headquartered in Milan, dual-listed on Milan and Oslo, chasing an investment-grade rating, is a re-rating candidate in a sector the European market had written down for a decade. If the IG target lands, the multiple moves, and that pulls capital back toward European energy-services equity that has been starved of it. Third, watch the offshore-wind read. TotalEnergies’ sharpest argument was that this concentration extends into wind installation and decommissioning, the next decade of demand, not just today’s oil work. If high-spec installation capacity is scarce for SURF, it is scarce for foundations and cable-lay too, and the same firm now sits on both. How a desk plays it Three expressions, and where each is wrong. The cleanest long is the combined equity into the investment-grade re-rating, the one genuine multiple lever in a sector the market wrote off, played on the dual Milan and Oslo listing. That long is wrong if the agencies stall on IG through the integration spend, so the stop is a ratings outlook that does not move toward IG by the FY27 results. The relative-value line is long Saipem7 against TechnipFMC, which sits at around 17% and owns the trees and SPS hardware the merged firm does not touch. That pair is wrong if Petrobras unbundles fast and rewards the equipment specialists over the installers, which is the same signal that breaks the pricing-power thesis. And the discrete credit item is the €500m 2029 convertible: completion triggers change-of-control, so model the refinancing or the dilution as an event, not a footnote. None of this is a view on the close. The close is not in doubt. It is a view on whether unconditional clearance turns into margin, and the first data point lands on the next award. The watchlist Six things settle this argument, and none of them is a press release. SURF award concentration and price per comparable scope on the next Petrobras rounds, dated to the award month. Whether Allseas wins a third Brazilian package or stalls at two. Whether Petrobras moves a real share of pipelay to direct charter. Vessel utilisation across the combined fleet, because that is the swing factor on the €300m and the tell on whether capacity is being kept tight. The ratings-agency timing on investment grade through the integration spend. And any appeal of the CADE clearance to the Tribunal. The deal is closing. The read on whether unconditional clearance actually shifted pricing power is a 2026-2028 data series, and it starts with the very next award. The call for the desk: model the subsea entity, not the holdco. Treat the 44% and the eight-of-twelve vessel share as the reason pricing power tilted toward the contractor, and treat the entrant, the unbundling and Petrobras’ own scale as the reason it did not become a blank cheque. The brake the operators wanted is off. Whether the car actually accelerates is the thing to watch, award by award. Charts rendered Engine A (native SVG). The board-format desk pack and the editorial cartoon ship as separate assets. Best Regards, Energy Pipeline is educational market commentary, not investment advice. The author is Managing Director of Germini Energy, a crude and refined products brokerage, and may have commercial interests in markets discussed. This publication does not comment on cargoes, counterparties, or transactions the firm is actively working. Market-share figures are a bottom-up reconstruction from public contract registers and are illustrative, not audited. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

10. 

    22 Jun 2026

    The Bill for the Boom

    For two years the AI power story lived in slide decks. Gigawatts by 2030. Curves that bent toward the sky. It was a forecast, and forecasts are easy to ignore when they arrive as a chart. That period is over. The number stopped being abstract the month it showed up as a line item on a household electricity bill in Ohio, in Virginia, in Texas. The boom found the kitchen table. Here is the plain version. Building the machines that run artificial intelligence requires an amount of firm, around-the-clock power that the American grid was not built to hand over on this timeline. Somebody has to pay to close that gap. The fight in 2026 is over who. The hyperscaler with a trillion-dollar balance sheet, or the family three counties over who never asked for a data center and cannot move away from one. This is not a technology story anymore. It is a cost-allocation fight, and the household is losing it. The demand is real, and it is fast Start with the load, because everything else follows from it. US data centers drew about 31 gigawatts of power in 2025. Goldman Sachs Research puts that at 41 gigawatts this year and 66 next year. That is more than a doubling in two years. To make the number physical: one gigawatt is roughly a large nuclear unit, or the steady draw of around 750,000 homes. The grid is being asked to find dozens of those, fast, in specific places. Speed is the part that breaks the system. A model gets retrained in months. A campus gets announced in a press release the same week it is conceived. Power does not work that way. A transmission line takes years. A substation takes years. A new generating unit takes longer. The demand side of this equation moves at the speed of software. The supply side moves at the speed of concrete, copper, and steel. That mismatch is the whole story, and no amount of capital erases it on command. The electron obeys physics, not a product launch calendar. The bill is already moving When demand jumps and supply cannot follow, price does the adjusting. That is not ideology. It is the merit order. And the adjustment has started. US utilities asked regulators for more than 29 billion dollars in rate increases in the first half of 2025 alone, double what they sought in the same window of 2024. Some of that is wildfire hardening and aging wires. A growing share is the cost of wiring up loads that did not exist when the system was planned. The Federal Reserve Bank of Dallas, not a body given to drama, modeled what doubling data-center demand does to wholesale power prices. Its range runs to as much as 50 percent higher, with selected months above 40 percent in the harder scenarios. Virginia, the place with more data centers per square mile than anywhere on earth, is already the cautionary tale its neighbors quote. The pattern is simple and it repeats. The data center signs a long contract at a price it can absorb. The grid spends billions to serve it. The cost of those billions gets spread across everyone on the system through the rate base. The hyperscaler gets the compute. The ratepayer gets the bill. That is the quiet transfer at the center of this. A private commercial decision, socialized across a captive public. No one voted for it. It happens at the public utility commission, in dockets written in a language designed to be skimmed past. The supply side cannot scale at the speed of hype Now the part I know in my bones, because it is an operations problem, not a finance one. You cannot conjure firm power. You build it, and building it runs into hard physical bottlenecks that capital cannot bid away. Take the gas turbine, the workhorse the industry reaches for first when it needs dispatchable power quickly. The wait to procure a large one has stretched from the old one-to-three years to five, and in cases seven. Siemens Energy is carrying the largest order backlog in its history, around 136 billion euros. The cost to build a new gas plant in the United States has jumped on the order of two-thirds. This is the supply side telling you, in the only language it has, that it cannot keep pace. When the queue for the machine that makes the power is longer than the contract to sell the power, you do not have a financing gap. You have a physics gap. So the industry reached for the other firm, clean option, and this is where 2026 gets genuinely interesting. Nuclear came back into fashion not for the climate brochure but because it is the only round-the-clock, carbon-free megawatt that a hyperscaler can point a data center at. Microsoft signed a roughly 16 billion dollar deal to restart Three Mile Island Unit 1. Amazon put money into X-energy and a multi-billion campus alongside it. Google backed Kairos Power. Meta lined up the largest book of all, more than six gigawatts across several developers. Add it up and you get about 9.8 gigawatts of announced nuclear capacity tied to AI. Read the dates, not the dollar signs. The Three Mile Island restart targets 2027. The small modular reactors that everyone is excited about mostly target 2030 and beyond. The demand curve in Chart 1 doubles by next year. The clean firm supply meant to meet it arrives, at the earliest, the year after, and mostly at the end of the decade. There is a gap of several years between when the power is needed and when the good answer shows up. Something has to fill it. In the near term, that something is the existing grid, more gas where it can be built, and the ratepayer’s bill absorbing the strain. What an operator sees that the forecast misses I have spent a career on the delivery side of energy, where the press release meets the P50 outcome and usually loses. So let me say what the buildout decks tend to leave out. First, “announced” is not “energized.” A large share of the data-center capacity penciled for this year will slip. Roughly half of US capacity planned for 2026 is already delayed or cancelled, held up not by money but by transformers, switchgear, interconnection queues, and a grid that takes up to five years to say yes. The capital is committed. The megawatts are not. That gap between the announcement and the operating asset is the most honest number in the whole sector, and it is the one least quoted. Second, firm power is the bottleneck, not generation in the abstract. A gigawatt of solar is not a gigawatt a data center can run on, because the data center runs at 3 a.m. in still air. What these loads need is power that is there every hour, and that is the scarcest, slowest, most expensive thing to build. The whole argument compresses to one sentence: the constraint is firm electrons, and firm electrons do not scale at the speed of a model release. Third, this lands on the household before it lands on the hyperscaler, and the politics will follow the bill. There is already a public mood turning against data centers in places that a year ago welcomed them. When the power bill rises and the new jobs turn out to be a few dozen technicians behind a fence, the welcome cools. Watch the rate cases. Watch which states start forcing large loads to pay their own way through special tariffs rather than spreading the cost. That fight, utility commission by utility commission, decides who actually pays for the boom. It is duller than a chip launch and it matters more. The thing to hold onto The demand is genuine. The capital is genuine. The intelligence being built may well be worth every watt. None of that is the question. The question is who carries the cost of closing the gap between a demand curve that doubles in two years and a supply chain that measures delivery in half-decades. Right now the answer is being decided quietly, in dockets, and the household is paying for a boom it was never asked to join. The AI story has been told as compute, chips, and models. The real story underneath it is older and harder. It is about firm power, long lead times, and who pays when the two do not line up. That bill is coming due. It has a name on it. For now, it is not the one you would expect. Best Regards,Felipe Vigne Germini Energy Pipeline is educational market commentary, not investment advice. The author is the Managing Director of Germini Energy, a crude and refined products brokerage, and may have commercial interests in markets discussed. This publication does not comment on cargoes, counterparties, or transactions the firm is actively working. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

11. 

    6 Jun 2026

    The Price Cap Is Dead. Washington Is Pretending Otherwise.

    Russian crude is trading at a premium to Brent. Read that line twice, because it ends an argument Washington and Brussels have been having with themselves since 2022. The entire Western sanctions architecture rested on one number. The discount. The G7 price cap, the European push toward a 44 dollar ceiling, the shadow fleet math, the insurance bans, every piece of it assumed Russia would sell cheap because it had nowhere else to go. The discount was the proof the policy worked. It was the scoreboard. As of this week, that scoreboard reads in Moscow’s favor. The Atlantic Council’s Energy Sanctions Dashboard, the one half my inbox forwarded last week, puts Russian grades at premiums of up to ten dollars over ICE Brent. It mentions this almost in passing, three screens down, wrapped in cautious think-tank prose. I will not be that polite. The price cap is dead. The only open question is whether anyone in Treasury will say so out loud before June 17. Chart 1. The spread that justified the entire price cap has flipped sign. What actually happened, stripped of drama Here is the sequence. In late March, the Strait of Hormuz effectively closed. Roughly a fifth of the world’s seaborne oil moves through that gap of water, and for the first time the IEA reached for a phrase it had never used before: the largest supply disruption in the history of the oil market. A ceasefire was announced on April 8. Traffic never returned to where it was. Tankers still treat the Gulf as a war zone because, functionally, it remains one. That is the supply shock. Now the policy response. Washington, staring at a Gulf it could no longer count on, did the one thing it swore it never would. It opened the door to Russian barrels. Bessent issued narrow waivers, let them lapse, reissued on April 17, then extended again. The latest extension runs to June 17. On June 2, Rubio sat in front of the Senate Foreign Relations Committee and said the administration wants to kill the waivers as soon as we possibly can. Two senior officials, pointing in opposite directions, on the same policy, in the same month. That is not strategy. That is a government improvising in public. Brent sits near 92 dollars as I write, down roughly 20 percent from the spring panic peak, holding in a 90 to 100 band while the market waits to see whether the ceasefire holds. The calm is borrowed. The structure underneath it has changed in ways that will outlast the headlines. Chart 2. One sanction, four reversals, ninety days, and a cliff on June 17. The scale nobody wants on the table Three sanctioned regimes, Russia, Iran, and Venezuela, accounted for almost 14 percent of global crude shipping in 2025. Russia alone moved 8.1 percent. Iran added 3.9. Venezuela, even with Maduro’s capture in January and the US blockade, still managed 1.8. You cannot fine one barrel in seven out of the market and expect the price not to notice. For three years the answer to that math was the shadow fleet, the discount, and a buyer in Beijing willing to pocket the spread. At peak discount in 2025, China was saving up to 28.8 million dollars a day buying sanctioned crude. That is not evasion at the margin. That is a parallel market with its own price, its own fleet, and its own logic. The Hormuz shock broke the discount. When Gulf barrels vanished, the world stopped treating Russian crude as contraband to buy cheap and started treating it as supply to secure at any price. The discount became a premium. The shadow fleet stopped being a discount mechanism and became a delivery mechanism. Same ships. Opposite economics. Chart 3. Sanctioned crude was roughly 14 percent of 2025 shipping flows. The chain of causality, one link at a time Walk it with me. This is where the consequences compound. 1. Hormuz closes. Roughly 20 percent of seaborne oil is suddenly unreliable. Asian refiners who built their diet around Gulf grades face a hole they cannot fill from inventory. 2. The United States waives its own sanctions on Russian crude. The instrument built to starve Moscow becomes the instrument that feeds Asia. 3. Russian export volumes surge. March imports of Russian crude rose about 41 percent over February, to levels not seen even before the 2022 invasion. Since the waivers, Russia has put roughly 300 million barrels back into the international market. 4. The buyer base widens. India returns to pre-sanction volumes and is openly lobbying to extend the waiver. Southeast Asia, a region that imported up to 96 percent of its crude from the Middle East, shows up as a brand new customer. The Philippines declared a national energy emergency. Indonesia went to Moscow and came home a buyer. Thailand, Vietnam, Malaysia, Sri Lanka are all in the queue. 5. New dependencies set like concrete. This is the link nobody in Washington wants to look at. A refinery that reconfigures for Russian crude does not switch back the morning a waiver expires. A government that signed an emergency supply deal with Moscow does not tear it up on a US press release. The waiver was sold as temporary relief. It is building permanent plumbing. 6. The discount inverts. With demand chasing a shrinking pool of available barrels, sanctioned exporters stop competing on price and start collecting a premium. Russia, the most sanctioned major exporter on earth, becomes the best-positioned seller in the market. 7. The enforcement tool loses its teeth. You cannot run maximum pressure on Iran with one hand while waving Russian cargoes through with the other and expect anyone to believe the threat. China noticed. For the first time, Beijing deployed its prohibition order, instructing its own companies to ignore US secondary sanctions outright. The mask of quiet compliance is off. Seven links. Each followed logically from the last. None of them were the plan. The enforcement problem, in one paragraph None of this works without the ships. Transshipment, the cargo-to-cargo transfer at sea that launders a sanctioned barrel into a clean one, and the shadow fleet that carries it, vessels with opaque ownership that switch off their transponders or spoof their position, remain the load-bearing wall of the whole evasion structure. The West keeps sanctioning individual tankers and individual Chinese teapot refiners. FinCEN issued a fresh advisory to banks on the red flags of Iranian oil smuggling. Treasury rebranded the Iran campaign as Operation Economic Fury and pointed secondary sanctions at intermediaries in China, the Emirates, Hong Kong, Iraq, and Oman. All of it is real, and none of it has shrunk the fleet. You cannot out-designate a problem that grows a new shell company every week. Enforcement that is not coordinated across the G7 and is not aimed at the financial plumbing is theater with a press release attached. FOR PAID SUBSCRIBERS This is the part that matters for anyone with a barrel, a refinery, or a budget exposed to diesel. Below the wall: three scenarios for the June 17 waiver cliff and how each prices through to Brent, Urals, and Brazilian diesel. The Global Economic Ripple, traced through Asia, Europe, and the one region the Atlantic Council dashboard ignores entirely, Latin America, where the realignment is quietly minting winners. And the single move I would make this week sitting on physical length or short diesel. Who this serves: the operator deciding whether to hedge June and July diesel, the trader weighing Urals length, and the executive who has to explain to a board why a temporary waiver is a permanent risk. Companion deck: 12 slides, including the discount-to-premium inversion mapped against the Hormuz timeline, and the Latin American barrel map most desks have not drawn yet. Subscribe to read the full analysis. Free readers get The Signal each Monday. Paid and Founding readers get The Flow and The Horizon, the scenarios, and the deck. Alternative scenarios for June 17 Scenario A. The waiver dies on June 17. Rubio gets his wish. Treasury lets it lapse and returns to undiluted maximum pressure. On paper, credibility is restored. In practice, you have pulled 300 million barrels of relief out of a market still short Gulf crude. Brent spikes back toward the spring highs. India and Southeast Asia, told they cannot legally buy Russian, buy it anyway, because a fuel shortage is a domestic crisis and a US fine is a line item. Indian officials have already said as much. Expiry does not stop the flows. It moves them back into the dark, rebuilds the discount for China, and hands Beijing the spread again. Sanctions look tough and accomplish less. Scenario B. The waiver is extended quietly, again. The most likely outcome, and the most corrosive. Washington keeps issuing thirty and sixty day extensions because the alternative is an allied recession. Each extension normalizes Russian crude a little more. Urals holds its premium. The price cap becomes a footnote nobody bothers to repeal. Moscow funds its budget at market prices while the West congratulates itself on a cap that no longer caps anything. Scenario C. Hormuz reopens faster than expected. The ceasefire holds, Gulf volumes recover over the summer, and the supply panic drains out of the price. This is the bullish case for sanctions and the bearish case for Brent. Russian premiums collapse back to discounts as Gulf barrels return. But even here, the Southeast Asian relationships do not vanish. Indonesia and the Philippines learned a lesson about single-source dependence that no reopening unteaches. The flows shrink. The plumbing stays. My weighting: B is the base case, A is the political tail risk markets are underpricing for June, C is the hope every importing finance minister is praying for and none can bank on. Chart 4. Every option trades market stability against sanctions credibility. Global economic ripple Asia. The epicenter. The region went from comfortable Gulf dependence to a frantic supply hunt in ninety days. Diesel and gasoline prices in the Philippines doubled in the first month of the crisis. Thailand, Vietnam, and Malaysia are exposed at 70 percent import reliance or worse. The structural shift is that Asian energy security now runs partly through Moscow, and that is a strategic gift to Russia no battlefield delivered. China, meanwhile, sat on 169 million barrels of fresh storage it built in 2025 and 2026 and is playing every side: stockpiler, swing buyer, and now open defier of US secondary sanctions. Watch the Chinese teapot refineries. They are the pressure point Treasury keeps designating and never quite closes. Chart 5. China holds the storage, the volume, and now the willingness to defy. Europe. The continent that built the price cap is watching it die and saying very little. The EU pushed toward a 44 dollar ceiling precisely when the market moved to a Russian premium, which means the cap is now a number with no buyers and no relevance. Brussels has two bad options: enforce a cap the market ignores and look impotent, or quietly let it lapse and admit the pressure is gone. Either way, the lesson is the one Europe should have learned in 2022. A sanction is a tool of pressure only when the sanctioned party has somewhere worse to go than you. Russia now has Asia. Latin America. Here is the region the dashboard does not mention, and the one I watch most closely from Sao Paulo. The realignment is quietly minting winners across the Atlantic Basin. Every barrel of Gulf crude that does not reach Asia is a barrel someone else has to supply, and non-sanctioned medium and sweet grades gain. Brazilian pre-salt crude, already pointed at China, becomes more valuable simply by being clean, liftable, and politically boring. Guyana’s output looks better by the day. Argentina’s Vaca Muerta exports catch a tailwind. I have brokered enough of these flows to know the arbitrage does not care about ideology, only about delivered cost. For a decade the diesel that clears the Brazilian market has been a contest between US Gulf Coast cargoes and, more recently, Russian barrels arriving through long, deniable supply chains. Close Hormuz, scramble the Russian diesel pool, and distort the Gulf Coast arb at the same time, and you remove the two reference points that Brazilian importers price against. The result is not a clean shortage. It is a wider, more volatile import parity that Petrobras either passes through to the pump or absorbs onto its own balance sheet for political cover. Either choice has a constituency that screams. The other side of the ledger is diesel. Brazil is a structural diesel importer, and the global diesel complex is the most exposed product in this whole mess. With Russian diesel flows scrambled and the US Gulf arbitrage distorted, Petrobras faces the old, ugly choice between import parity pricing and political pressure to eat the spread. Watch the Brazilian pump. It is where a strait in the Persian Gulf turns into a line item in a Belo Horizonte trucking budget. And Venezuela, with Maduro gone since January and general licenses redirecting crude north, is being slowly reabsorbed into a US-facing supply orbit. The heavy barrel map of the hemisphere is being redrawn while everyone stares at Hormuz. Chart 6. The plumbing being built now does not reverse on a press release. The strategic read If you want the one sentence I would put in front of a board: the waiver is not the story, the dependency is. Markets are trading June 17 as a binary, expiry or extension, spike or relief. That is the wrong frame. Whatever Treasury decides on the 17th, the structural fact is already set. A meaningful slice of Asian and emerging-market demand has been rewired to run through Russian crude, and that wiring does not reverse on schedule. The premium is the proof. For the operator, that means treating Russian length as a market position, not a sanctions risk, and hedging diesel exposure into the third quarter rather than betting on a clean reopening. For the policymaker, it means admitting that a sanction you cannot afford to enforce is not a sanction. It is a bluff with a deadline. And for anyone in Brazil or the wider Atlantic Basin, the next eighteen months hand non-sanctioned producers a structural premium they did nothing to earn, and hand diesel importers a bill they cannot avoid. The Atlantic Council ends its dashboard asking how Washington can stabilize markets without weakening pressure on Russia and Iran. After this spring, that question answers itself. It cannot. It already chose. The barrels chose for it. Best Regards, Felipe Vigne Germini is an energy executive, operator, and dealmaker with 25 years across the oil and gas value chain, from deepwater operations to physical crude and product trading in Brazil and Latin America. He writes Energy Pipeline, an independent newsletter on energy markets, geopolitics, and the business of moving barrels. ENERGY Pipeline by Felipe Germini is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

12. 

    5 Jun 2026

    OPEC+ Sunday: A Floor Without Barrels

    This is a free preview of a paid episode. To hear more, visit fgermini.substack.com (https://fgermini.substack.com?utm_medium=podcast&utm_campaign=CTA_7) ENERGY PIPELINE by Felipe Vigne Germini THE HORIZON  |  Friday Edition  |  June 5, 2026 This Sunday the OPEC+ ministers sit down in a room with a missing chair. The Emiratis walked out on the first of May. Fifty years of membership, closed in a press release. The group that meets this weekend is not the group that wrote the rules a year ago, and the market is still pricing it as if nothing structural has shifted. Dated Brent finished the week near 96 dollars. WTI a few dollars under. Glance at the tape and you would call this an ordinary tight market. It is not. Think of OPEC+ this weekend as a smoke detector with a dead battery. The price is held up by a war premium, not by discipline, and the cartel that is supposed to manage the floor has quietly lost the one thing that made it credible. The barrels. Start with the arithmetic, because the arithmetic is the whole story. Saudi Arabia’s June allocation puts its required production at 10.291 million barrels a day. Its actual output is closer to 7.25 million. That is a gap of roughly three million barrels a day between the quota on paper and the oil that physically leaves the ground and reaches a buyer. This is not a voluntary cut announced from Vienna. It is an involuntary cut imposed by a closed Strait of Hormuz, where traffic has fallen to a trickle, on recent vessel counts somewhere around five to ten percent of normal. The Saudis are not choosing restraint. They are stranded. The shock absorber that isn’t For twenty years the market slept at night because of one number. Spare capacity. The few million barrels that Gulf producers could bring online inside thirty days. That buffer was the shock absorber. It is the reason a refinery manager in Rotterdam or a finance minister in Delhi could assume any supply shock had a ceiling. Somebody in the Gulf would open the taps. That assumption is now wrong, and the chart below shows why. The buffer was never just a Saudi favor to the world. It was the asset that let everyone else under-invest in resilience. Importers ran lean inventories. Refiners hedged thin. Airlines and shipping lines wrote fuel budgets on the quiet assumption that the Gulf would always backstop a spike. Strip the backstop out and every one of those decisions has to be re-underwritten at the same time. That is the part the screen has not priced. Chart 1. The deliverable buffer, the oil that could actually reach the market next month, is a rounding error. Before the UAE left, OPEC+ claimed a spare buffer near six million barrels a day. The Emiratis took 1.54 million of that with them, the second largest single share after Saudi Arabia. What remains sits overwhelmingly inside the Gulf, behind a strait that almost nothing is transiting. Spare capacity that cannot be shipped is not spare capacity. It is a slide in a deck. The deliverable buffer is a rounding error.

13. 

    2 Jun 2026

    The market has the price about right. It has the shape badly wrong.

    Every bank has the same Brent number for 2026. Ninety-six dollars. Or ninety-five. Or a hundred. After a war premium that pushed crude to a hundred and seventeen in April, after the Gulf shut in ten million barrels a day, the sharpest desks on the Street all landed inside a fifteen-dollar band. Everyone calls that confidence. I call it a crowded boat. Here is the part the consensus gets backwards. The number is probably about right. Ninety-six dollars is a defensible average. The danger was never the level. It is the shape. The forward curve is paying a thirty-dollar premium to deliver now versus a year out. That is the market charging real money to carry a position through the year, and almost nobody budgeting around the headline has priced it. Right on the number, wrong on the structure. And if you are an importer, a refiner, or a finance minister hedging the flat price and calling it risk management, you are standing on exactly the same rail as everyone else. There are three ways this cluster breaks. Only one of them is the orderly fade everyone penciled in. The other two do not show up in the average at all. They show up in the structure, which is exactly where most people forgot to hedge. Here is the part the sell side will not say out loud. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

14. 

    29 May 2026

    EP PODCAST: The Freight Market Already. Priced the Deal. The Strait Has Not Been Told.

    The Platts assessment for USGC-Brazil 38kt MR clean products closed at $40.70/t on May 26. Six weeks earlier, on April 14, the same route was at $137.36/t. The freight market has returned the entire Hormuz war premium. The Worldscale rate is back to WS 200, exactly where it was on February 2, before the Iran war started. The market priced the 60-day ceasefire framework, the de-mining agreement, the anticipated reopening. It ran on the headline — before a single mine was cleared. Here is where it gets complicated. P&I war risk insurance, withdrawn effective March 5, is still suspended. No commercial VLCC has transited the strait. The 68 LR2 tankers that switched from clean products to dirty crude trading during the crisis are still in the dirty market. The physical supply side of the clean tanker market has not confirmed what the Worldscale rate is pricing. WS 200 today is a bet that the deal executes on schedule. The base case says it doesn't. And the base case puts freight 40% above today's print within 30 days. Three scenarios, four variables to watch, and what Petrobras's June contracting behavior signals about which way this goes: full Horizon analysis on Substack, paywalled. Link in the first comment. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

15. 

    27 May 2026

    EP PODCAST: The Insurance Geometry That Replaced the Strait

    Tuesday night, US Central Command struck two IRGC mine-laying boats in the Strait of Hormuz and a SAM site at Bandar Abbas. Brent gave back two-thirds of its reopening rally by Wednesday open. The Lloyd's market did not move at all. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

16. 

    26 May 2026

    EP PODCAST: The War Premium Refuses to Die

    The War Premium Refuses to Die The trade is the crack, not the flat price. THE FLOW  |  WEEKLY ENERGY PIPELINE RECAP  |  MAY 18-22, 2026  |  ISSUE W21/26  |  FELIPE GERMINI Most macro coverage this quarter is anchored on OPEC+ quotas and Chinese demand. The actual price discovery is happening inside Lloyd's of London. Since late February, war-risk premiums on Strait of Hormuz transits have moved from 0.15% of hull value to 5.5%, translating into USD 3M to 8M of insurance cost per VLCC voyage. That is roughly $3 per barrel of pure premium before freight, before lightering, before P&I. Brent printed $105.10 on Friday and still closed the week down 4.3%. The market is pricing the reopening of Hormuz before the diplomatic paperwork is signed. The crack complex has not unwound in tandem. That gap, between flat price selling off on hope and crack spreads holding the wartime premium, is the trade. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit fgermini.substack.com/subscribe (https://fgermini.substack.com/subscribe?utm_medium=podcast&utm_campaign=CTA_2)

Contact ENERGY Pipeline by Felipe Germini Podcast

Guest appearances

Does not typically book guests

Based on episode analysis; this does not confirm that the show is currently accepting guests.