CV Research(@CVResearch_):$BE is worth about $85 billion. $INIO is worth $14.4 billion. This year they will book almost the same revenue. $INIO guides to $3.8 to $3.9 billion in 2026 revenue. $BE guides to $3.9 to $4.2 billion. INNIO shipped 3.4 gigawatts of engines last year. Bloom ended 2025 with about 1 gigawatt of annual factory capacity. INNIO is building toward 10 gigawatts a year by 2030. An AI campus needs power on a specific date and the grid cannot promise one. In the busiest regions, connecting a large new load takes five to seven years. Sell-side estimates value getting power one year earlier at about $4.50 per watt. INNIO's engines sell for $0.45 to $0.70 per watt. The buyer is paying for the date, not the hardware. The J624 is a 5 megawatt gas engine that gets trucked in on a trailer and takes load in about 15 seconds. Nothing larger takes load that fast. The demand is already on paper. Backlog went from $1.0 billion at the end of 2024 to $6.6 billion in June. Customers have prepaid $1.4 billion. 2026 and 2027 are sold out and reservations reach into 2030 and 2031. We traced about 8 gigawatts of Jenbacher engines to named sites. $AMZN specified 280 J620s in a Pennsylvania permit. $ORCL and OpenAI's Frontier campus runs on Jenbachers through VoltaGrid. Liberty Energy filed the first real price in an 8-K: $493 million for 100 J624s. Then there is the part nobody prices. Services were 48% of 2025 revenue and about 65% of segment profit. A data center engine runs about 8,000 hours a year against 2,000 to 4,000 on the legacy fleet, and the parts wear by the hour. The first wave hits major overhauls from 2029. So why is it $19.20, 29% below the $27 IPO and 55% below the June high? Because of who is selling. The IPO was 100% secondary. The sponsors still own 86% and only about 14% of the company trades. One quarter of tariff hit margins did the rest. The lockup expires November 30. Now run the math. At $19.20 the market is pricing capacity that stops at about 5.7 gigawatts, short of the Austrian expansion already under construction, with the new Wisconsin line adding nothing. Our base case is $27, about 40% above today. Bear is $14.40 and bull is $38. The risks are real. The Texas permit pause. A sponsor with 646 million shares to sell. VoltaGrid at 41% of 2025 orders. All of it is in the report, along with what would change our mind. The dates that decide it: October 19 TCEQ's report, October 29 Enerflex earnings, mid November INNIO's Q3, November 30 the lockup, and around December 10 when ERCOT finishes its review. The grid will get there eventually. The data center cannot wait for it. INNIO sells the time in between.

2026.10.09 15:34

$BE is worth about $85 billion. $INIO is worth $14.4 billion. This year they will book almost the same revenue. $INIO guides to $3.8 to $3.9 billion in 2026 revenue. $BE guides to $3.9 to $4.2 billion. INNIO shipped 3.4 gigawatts of engines last year. Bloom ended 2025 with about 1 gigawatt of annual factory capacity. INNIO is building toward 10 gigawatts a year by 2030. An AI campus needs power on a specific date and the grid cannot promise one. In the busiest regions, connecting a large new load takes five to seven years. Sell-side estimates value getting power one year earlier at about $4.50 per watt. INNIO's engines sell for $0.45 to $0.70 per watt. The buyer is paying for the date, not the hardware. The J624 is a 5 megawatt gas engine that gets trucked in on a trailer and takes load in about 15 seconds. Nothing larger takes load that fast. The demand is already on paper. Backlog went from $1.0 billion at the end of 2024 to $6.6 billion in June. Customers have prepaid $1.4 billion. 2026 and 2027 are sold out and reservations reach into 2030 and 2031. We traced about 8 gigawatts of Jenbacher engines to named sites. $AMZN specified 280 J620s in a Pennsylvania permit. $ORCL and OpenAI's Frontier campus runs on Jenbachers through VoltaGrid. Liberty Energy filed the first real price in an 8-K: $493 million for 100 J624s. Then there is the part nobody prices. Services were 48% of 2025 revenue and about 65% of segment profit. A data center engine runs about 8,000 hours a year against 2,000 to 4,000 on the legacy fleet, and the parts wear by the hour. The first wave hits major overhauls from 2029. So why is it $19.20, 29% below the $27 IPO and 55% below the June high? Because of who is selling. The IPO was 100% secondary. The sponsors still own 86% and only about 14% of the company trades. One quarter of tariff hit margins did the rest. The lockup expires November 30. Now run the math. At $19.20 the market is pricing capacity that stops at about 5.7 gigawatts, short of the Austrian expansion already under construction, with the new Wisconsin line adding nothing. Our base case is $27, about 40% above today. Bear is $14.40 and bull is $38. The risks are real. The Texas permit pause. A sponsor with 646 million shares to sell. VoltaGrid at 41% of 2025 orders. All of it is in the report, along with what would change our mind. The dates that decide it: October 19 TCEQ's report, October 29 Enerflex earnings, mid November INNIO's Q3, November 30 the lockup, and around December 10 when ERCOT finishes its review. The grid will get there eventually. The data center cannot wait for it. INNIO sells the time in between.