Pershing Square Challenge 2026 runner-ups on Baker Hughes $BKR

Summary

• Columbia’s Persian Square Challenge runner-ups (Carl, Cam, Greer) pitch Baker Hughes (~$65) as a misunderstood hybrid: half legacy oilfield services, half industrial energy technology (IET) — gas turbines, LNG equipment, and long-dated service contracts. IET has gone from 37% of the mix in 2020 to roughly 50/50 in 2025; the team asks whether the market is pricing a 2029 backlog-conversion story, while they argue “this is a 2030 and beyond story” — a “long-term compounder.”
• The bear case is that Baker Hughes has topped near $65 three times (2007 peak oil, 2015 shale peak, now) — the team’s answer is this turbine cycle is structurally different. Prior booms faded for distinct reasons (“Enron happened, market popped, and that was it for gas turbines”; 2015 renewables subsidies), while today AI, gigawatt-scale utility orders, coal retirements (Youngstown replaced one coal plant with 16 utility-scale gas turbines), and battery storage that still needs generation when the sun is not shining all converge. Cam: “AI and data centers can almost be like a little bit of a distraction” — onshoring and electrification drive demand regardless.
• Even if AI demand vanished, the team argues turbine tiers backfill each other because components share supply-chain resources. Sub-20MW (data centers, Baker Hughes’s NovaLT line), sub-100MW (LNG/industrial), and >100MW grid-scale compete for similar parts with different end customers, so weak small-turbine demand is “almost taken over immediately” by mid/large — “it really doesn’t change the math.”
• The margin engine is services: 10-year-plus agreements at roughly double equipment margins, with recurring revenue going from under a third today to over 35% by 2030. The team says Baker Hughes will not prioritize Google over customers it has served for 70-plus years — “you’re going to wait 36 months while I deliver to Google first?” — while competing OEMs that grabbed margin now face “very unhappy” customers. Andrew pushed the mercenary counter-case but conceded restraint is a margin of safety both ways.
• Valuation is deliberately a flat-multiple story: ~14x EBITDA today, sum-of-parts at 14.5x 2028 — the upside is IET growth plus the Chart acquisition. Chart, agreed to be bought for roughly $13.5B all cash (topping Flowserve plus a breakup fee, not yet closed), is modeled at ~$2B EBITDA / ~20% of 2028 EBITDA — Andrew says that implies Baker Hughes roughly doubles its money in four years, “5-10 billion dollars of value on a 70 billion EV company,” and invokes “winner’s curse, buyer beware.” Team comfort: decade-plus as Chart’s customer, conservative $325M cost-synergy guidance, and an LNG “one-stop shop.”
• The GE Aero Alliance handcuffs are coming off: the 2019 spin limited Baker Hughes’s turbine sales to oil & gas end markets only, but terms “loosened pretty significantly” in 2024 and Baker Hughes now sells NovaLT — wholly its own IP — into data centers. Carl’s read: Baker Hughes “gained a lot more from this merger than they lost,” and GE “probably isn’t too happy” about the value it spun away.
• Andrew’s governance flag: CEO Lorenzo owns ~$50M of stock but earns $22M/year, board ownership is skinny, and comp metrics are “never per share” — a worry when management agreed to an all-cash mega-deal in a cyclical industry. The team’s comfort came from disciplined divestitures alongside M&A and broadly positive expert calls: “current employees, past employees, customers, they love them,” though Cam noted he did not have a large sample size.

Deep dive

1. A two-headed energy company, chosen after PE-grade diligence

• Carl’s idea generation: energy as the “backbone of societal growth” — fleet electrification, data centers, rising electrons per capita — then backing into the name the market had not fully rewarded because of its legacy oilfield ties: “this sort of dichotomy of a stock that… slightly misunderstood or less understood than the market.”
• The diligence is the differentiator Andrew flags up front: 30+ expert calls plus the Western Turbine Users Conference in Long Beach — where Greer showed up with a broken arm (“Bill Ackman did sign the cast if anyone’s concerned”). Andrew’s framing: 30 expert calls is what you do “about to buy a multi-billion dollar company,” done here for a stock-pitch contest.
• The awkward validation: Baker Hughes hovered in the 40s through January-February when they picked it, then “rocket ship” to $65 by the time they pitched.

2. The market misses the momentum — a “2030 and beyond” story

• The business in a nutshell: oilfield services (equipment for extracting crude) and IET (equipment converting natural gas into electricity). IET was 37% of the mix in 2020, ~50/50 by 2025; Carl’s thesis is that the market “is not fully understanding kind of the magnitude of where this growth is” or how long the transformation runs.
• Carl’s sharper framing: “Is the market pricing this in as a 2029 growth and demand story?… this is a 2030 and beyond story” — the deck has a 3-year target, but the primary research points to “a long-term compounder.”
• Andrew raised the fear they heard throughout the semester — that this is “just another cycle.” Carl responds that this is “not just like 2010 or the ’90s.”

3. Andrew’s zoom-out vs. the convergence argument

• Andrew’s technical-analyst pushback: Baker Hughes has hit ~$65 three times — late 2007 (peak-oil fears, $100+ crude, pre-GFC) and ~2015 (shale-boom peak) — “it kind of looks like we’re paying for these businesses while everyone really likes them right now.”
• The team’s rebuttal: as the mix shifts, the question becomes a gas-turbine cycle, and prior turbine busts had distinct causes — early-2000s deregulation (“Enron happened, market popped, and that was it for gas turbines”), then 2015 renewables subsidies pulling capex. Now factors converge that “previously just didn’t exist”: AI/data centers, utilities globally ordering “in the gigawatt ranges,” coal retirements (Youngstown, Ohio swapped a huge coal plant for 16 utility-scale natural-gas turbines), and grid batteries that still need generation “when the sun isn’t shining over Texas.”
• Cam’s caution against the obvious narrative: “AI and data centers can almost be like a little bit of a distraction” — onshoring, electrification, and industrial buildout in the US and globally drive demand on their own.

4. The AI bear case, turbine tiers, and the loyalty-pricing debate

• Carl’s answer to “what if the AI bubble bursts”: three turbine categories — sub-20MW (data centers, Baker Hughes’s NovaLT line), sub-100MW (LNG/industrial), and >100MW grid-scale (utilities) — share similar components and compete for supply-chain resources with different end customers, so if small-turbine demand weakens it is “almost taken over immediately” by mid/large. If AI is gone tomorrow, “it really doesn’t change the math.”
• On pricing power: the team says Baker Hughes is not willing to tell customers of 70-plus years “you’re going to wait, you know, 36 months while I deliver to Google first” — a view validated with management and industry contacts. Greer’s point is that the real margin sits in long-term service agreements, so nobody rational makes “a quick buck on the equipment” at the cost of the relationship.
• Andrew’s skepticism — worth keeping: every company that claims to “do the right thing” ends up with customers saying “cut your prices”; maybe Baker Hughes “should just be more mercenary… Google is going to pay us $2,000 for this little light bulb.”
• Carl’s counter: unnamed competing OEMs did exactly that and their customers at the conference are “very unhappy”; the team “pushed really hard to validate” whether Baker Hughes was immune and concluded it avoids “short-term grabs of margin.” Andrew’s concession: the restraint is a margin of safety — realizable upside if things run hotter, stickier customers if they cool.

5. Flat-multiple valuation and the Chart double

• The valuation is intentionally a flat-multiple story: ~14x EBITDA at pitch, sum-of-parts implying 14.5x 2028 — upside comes from IET revenue growth and mix into service margins that industry conversations put at nearly double equipment margins, with recurring revenue under a third today going above 35% by 2030.
• Chart: announced mid-2025 at what Andrew recalls as ~$13.5B all cash — topping the Flowserve merger and paying a breakup fee, “about the highest price you can pay” — and not even closed yet. The deck shows roughly ~$2B of EBITDA (~20% of 2028 EBITDA); Andrew says it puts Chart’s 2028 value at something like $28B: “they’re going to double their money inside of 4 years… that’s 5-10 billion dollars of value on a 70 billion EV company.” His verdict on the risk: “winner’s curse, buyer beware.”
• The team’s comfort: Baker Hughes has been Chart’s client for a decade-plus, management guides to $325M of cost synergies and has historically been conservative in its projections while staying quiet on revenue synergies, and the deal makes Baker Hughes “the one-stop shop for all things in the LNG value chain.” Integration will run Chart as a separate business “for a while,” consistent with their culture of slow integration.

6. GE untangling, alignment questions, and the C3.ai lottery ticket

• Carl on the GE Aero Alliance: GE Aero makes the turbines and blade IP that both companies sell; per the 2019 spin, GE Vernova sells into everything except oil & gas while Baker Hughes was confined to oil & gas — restrictions “loosened pretty significantly” in 2024, and Baker Hughes now sells into data centers with NovaLT, “completely their own IP.” Net assessment: Baker Hughes “gained a lot more from this merger than they lost,” and GE “probably isn’t too happy… giving up so much of that value.”
• Andrew’s alignment flag: Lorenzo owns ~$50M of stock but earns $22M/year, board ownership is skinny, and proxy metrics like ROIC and FCF are “never per share” — a setup where management gets paid to buy and grow, “not necessarily when shareholders do the best.” Carl’s comfort is capital allocation: divestitures alongside acquisitions signal long-term platform focus, not empire-building. Expert calls were broadly positive, though Cam caveated that he lacked a large sample size — “current employees, past employees, customers, they love them” — with Lorenzo referenced on a first-name basis; 2027 marks two decades at the organization.
• The C3.ai oddity Andrew dug out of old proxies: Baker Hughes’s 2022 investment, with Lorenzo serving on C3.ai’s board, followed by Baker Hughes selling $100M+ of stock at good prices. The team’s answer shows the diligence depth — they spoke to the former CFO (a Columbia alum now leading gas-turbine and IET work at Baker Hughes). Carl floated that it might have been a talent-acquisition play and guessed it could relate to Baker Hughes’s predictive-analytics stack (“I believe it’s called Crescent”), which monitors turbines 24/7/365 across thousands of data points for runtime-based service agreements. Andrew’s parting shot on C3: “how much of the value of that company is just they have the ticker AI?”