Who’s making money on AI? So far, it’s like the gold rush when the makers of picks, shovels, and blue jeans made all the money. In the case of AI, the makers of picks and shovels include:
- Electrical infrastructure: ABB, Schneider Electric, Siemens, and Eaton
- HVAC: Trane, Carrier, Daikin, and Munters
- Compute: Nvidia, AMD, Intel, and Broadcom
- Standby generation: Caterpillar, Generac, and Cummins
- Engineering: Jacobs, AECOM, and EMCOR
Millionaires are also being made by untested speculative companies, especially in large-scale power generation (not backup), including fuel cells (Bloom Energy) and nuclear power (Oklo). Bloom Energy, for example, has a market cap of $83 billion, making it one of the nation’s largest 150 companies. Its market cap is greater than 3 M’s. Bloom has $2 billion in revenue and hasn’t turned a profit. For reference, mature companies have a market-cap-to-revenue ratio of 2-3, while Bloom’s is 40. Oklo’s ratio is near infinity because it has almost no revenue.
AI Boondoggles and Misguided Incentives
Elsewhere, users of AI are struggling with leveraging it. One notable example includes recent articles describing the AI boondoggle at Uber headquarters[1][2]. Uber’s 5,000 engineers in R&D burned through the company’s 2026 AI budget in four months, spending tens of millions of dollars very quickly. Power users reportedly spend $500 to $2,000 per month on the AI meter.
Why did that happen at Uber? You get more of what you reward and less of what you punish. Uber incentivized AI token consumption by ranking engineers on internal leaderboards for Anthropic’s Claude Code usage. The race was on to spend money rather than to produce anything useful, such as, uh, results. This is exactly the opposite direction we’re headed – looking for results, not activity, AI or otherwise. Forbes summarizes the unattractive results: Companies assume AI will pay for itself through higher productivity. The trouble is that AI usage and costs can increase five to twenty times faster than measurable output, making the return on investment far less obvious than hype suggests.
The Yahoo article notes that AI spend is exceeding the salaries and fringes of employees who are deploying them in some cases! “Uber’s chief operating officer this week went a step further, raising eyebrows by saying all this AI spending was showing no noticeable increase in productivity.” He’s probably not a heavy AI user. He’s in charge of results, not spinning hamster wheels and looking busy.
Isn’t AI Cool?
Artificial intelligence is a powerful tool to leverage. It can do some cool stuff, but so can a miniature poodle at a fraction of the cost and with no prompting. I remember my father, circa 1980, being fascinated by Intellivision technology, which brought video games to cathode ray tubes (an early version of a television, gen-zeer), before cable television even existed. See Figure 1.
This is the case today with AI. It can do amazing things, but is it faster and better for the task or workflow? Is it faster to create a workflow and get a return on investment for the time spent? That is the question for any use case. Clearly, the Uber COO answers “no“ to those questions for their R&D engineering team.
Anyway, I was deadly at several things in my day: ping pong (the real type, not on a video game) and Intellivision baseball are two of them. We could play a nine-inning game in 15 minutes and throw batters out at first base from the outfield.
Figure 1 1980 Edition of Artificial Intelligence – Intellivision Baseball
AI Disruption and Societal Reset
I have been in the audience of famous AI enthusiasts, including Peter Diamandis, host of the Moonshots podcast, and his frequent guest, Salim Ismail. I recently listened to another podcast on Diaries of a CEO with guests, Kevin O’Leary, Shark Tank, and a developer behind a 9 GW data center in Utah, and his curmudgeon counterpart, Cenk Uygur, a “left-wing political activist,“ per Wikipedia, and founder of the Young Turks news and commentary program. I’ve listened to A16Z podcasts featuring Elon Musk. They all believe that AI will do great things, including ending work as we know it, curing cancer, and more.
In the recent podcast with AI optimist O’Leary, doomsayer Uygur warned of massive civil unrest due to AI-driven unemployment. AI pioneer Roman Yampolskiy projects 99% unemployment by 2030. I guess it’s time to take that dream vacation on the eve of societal collapse and universal poverty, except for the tech bros who will own all the wealth.
All of the above argue that universal basic income (UBI) will become a necessity. Uh, yeah – if that happens, it will be a colossal mess. If there is no earning or ownership, there is no respect for property and wealth. History has shown us this with housing projects.
Get this, Elon Musk, formerly leading Trump’s department of government inefficiency, pronounced doggie, was quoted during his tenure in that position, “Universal HIGH INCOME via checks issued by the Federal government is the best way to deal with unemployment caused by AI.”
Immunity
Hold on, I say. First, developing AI cost-effectively is a major challenge, as I already described above. Until it can read minds and imagine things, deploying it for many custom workflows can take more work than doing it with a human brain, eyeballs, and fingers on a keyboard. Tails will wag dogs, as in the case of Uber. I put a stop to one such development effort at Michaels. Perfecting AI was throwing a pinch point, aka a bottleneck, into the development of something I needed months ago. Just do it!
Lastly, the AI tech bros noted above advise the little people to develop AI skills and agents on the edge of the organization. Why? Because of the immune response of core operators, especially middle managers. Chat gives me the following definition of an immune response in an organization: “An organization’s immune response to AI is the collection of cultural, procedural, and political forces that instinctively resist changes to established ways of working.”
The Wall Street Journal recently reported that the American Rebellion Against AI Is Gaining Steam. “The only thing growing faster than the artificial-intelligence industry may be Americans’ negative feelings about it.“ Hatred and loathing of the tech bros and their hyper-mega data centers is reaching a crescendo. The little people will have a say in this country, at least. Unfortunately, the Laobaixing in China do not. That is the risk Mr. O’Leary warns us about.
Buy Results, Not Hype
The bottom line for buyers of services in the energy efficiency and load management world: let pricing and results show who’s king. Don’t believe the hype.
[1]Forbes: Uber Burns Its 2026 AI Budget In Four Months On Claude Code
[2]Yahoo Finance: After the AI binge, companies balk at soaring bills
