Home
로그인
회원가입
학습 콘텐츠
Loading...
The $250 Billion Lie. Why Musk’s Elite Team All QUIT - Video học tiếng Anh
듣기 연습
듣기 연습
/
Video
/
The Infographics Show
/
The $250 Billion Lie. Why Musk’s Elite Team All QUIT
The $250 Billion Lie. Why Musk’s Elite Team All QUIT
학습 모드 선택:
자막 보기
단어 선택
단어 다시 쓰기
Highlight:
3000 Oxford Words
4000 IELTS Words
5000 Oxford Words
3000 Common Words
1000 TOEIC Words
5000 TOEFL Words
자막 (195)
0:00
In February 2026, Elon Musk made a $250B move that was supposed to change the
0:06
future forever. SpaceX merged with xAI, creating a $1.25 trillion super empire.
0:14
Rockets, satellites, and superintelligence… all under one roof. It looked unstoppable. But just
0:20
weeks later, something strange started happening behind the scenes. The elite team Musk personally
0:25
built began to disappear quietly, one by one. No headlines. No explanations. By March, they
0:31
were all gone. So what did they see that made them walk away from the most powerful company on Earth?
0:36
I’m Josh and on today’s episode of The Infographics Show, we’re explaining why Elon
0:40
Musk’s $250 billion AI empire is empty. Chapter 1: The False Advantage
0:46
For all his flaws, Elon Musk has always been ambitious… maybe too ambitious. Moon tourism,
0:52
Mars colonies, flying cars, high-speed underground tunnels… he promised a future
0:57
straight out of science fiction. Fans were captivated, imagining a world reshaped by his
1:02
vision. Some of his ideas have become real, but his boldest promises are still just pipe dreams.
1:07
Yet every failure, every setback, hasn’t slowed him down.
1:11
If anything, it only fueled his confidence. So, when he started his own AI business, xAI,
1:16
in 2023, he didn’t just want it to be another player in the industry. He wanted it to lead the
1:21
way and take machine learning to new heights. To do that, he needed to make the
1:25
best hardware in the business. That’s exactly what he set out to do.
1:28
In a nondescript corner of South Memphis sits a repurposed Electrolux plant. It’s
1:33
home to Colossus, the world’s largest supercomputer, designed to train xAI’s
1:38
chatbot Grok and other advanced AI systems. Setting up a data center normally takes 4 years.
1:44
Musk didn’t want to wait. He was told 2 years might be possible.
1:48
Still too slow. In the end, his team built Colossus
1:50
in just 122 days, shattering every expectation. When Colossus came online, it wasn’t just
1:57
powerful, it was a monster. Miles of fiber optics, 200,000 GPUs,
2:03
a maze of chips and circuits designed to devour data and train the smartest AI ever built. And
2:08
Musk wasn’t satisfied. He wanted more. Half a million GPUS… a million. He wanted to grow the
2:14
strongest supercomputer the world had ever known. And Colossus isn’t just powerful. It’s massive.
2:20
It’s got its own energy grid, battery banks, and gas turbines, sucking
2:25
hundreds of megawatts just to keep it alive. A system of this size and scale comes at a cost.
2:30
The hardware alone has been estimated at around $7 billion. xAI’s plans to add even more chips
2:36
could tack on tens of billions more, not to mention the insane cost of running Colossus 24/7.
2:43
But none of that seems to faze Elon Musk. His approach is simple. Throw enough chips
2:47
at Colossus, and it will beat his rivals - OpenAI, DeepMind, Microsoft - propelling xAI to the top of
2:53
the AI world. Every GPU, every dollar spent, is a bet on building the ultimate AI empire.
2:59
But there’s a fundamental flaw with that: raw silicon alone is not enough
3:03
to create the world’s best AI. You need people for that.
3:07
Human minds. Engineers, architects, designers, developers. Geniuses
3:12
who can channel all that raw computing firepower into world-changing innovations.
3:17
That’s why the other AI giants focus on people, not just hardware. They chase the brightest minds,
3:22
offering massive salaries and perks to lure top talent. Even the most powerful
3:27
supercomputer is useless without the humans who know how to make it work.
3:31
They’ve improved their technology, too, but haven’t gone all-in on GPUs
3:35
and processors in the same way that Musk has. xAI took a different approach - pouring money
3:39
and GPUs into building Colossus - yet failed to hold onto the brilliant minds
3:44
who built the company in the first place. Without them, all that power meant little.
3:48
So, while Musk might openly boast about Colossus’ computing power and record-setting tech specs,
3:53
behind the scenes, there’s a brutal reality. xAI may have all the gear, but has no
3:59
idea about how to use it effectively. And it didn’t take long for the
4:02
first warning signs to appear. Chapter 2: The Radioactive Confession
4:07
Grok launched to much fanfare and quickly became a core feature of the X platform.
4:11
But as new versions rolled out, something became clear: Elon’s “scary smart” AI wasn’t living up
4:17
to the hype. It wasn’t pulling ahead of rivals or outperforming ChatGPT or Google’s Gemini on
4:22
the tests that mattered. Far from it.
4:25
In fact, in numerous expert comparisons and benchmark breakdowns, Grok repeatedly fell short.
4:30
Its content creation was lacking, with the model often delivering bland and basic responses.
4:35
It was a far cry from the more nuanced and higher-quality content of its competitors.
4:39
Grok struggled to keep up with its rivals when it came to coding, leaving users frustrated. Studies
4:45
also showed it had a high error rate, generating incorrect citations and misleading information.
4:50
It wasn’t a bad model. But it wasn’t the be-all and end-all that Elon had been promising.
4:56
Then came something much more controversial. Grok seemed to go rogue.
5:00
In May 2025, it began ranting about white genocide in South Africa, even when asked
5:05
about completely unrelated topics. A couple of months later, reports emerged of Grok praising
5:10
Adolf Hitler and spreading antisemitic content. It went on to aggressively insult various world
5:15
leaders, like the Prime Minister of Poland, Donald Tusk, and spread false claims following
5:20
the assassination of Charlie Kirk. It was embroiled in an even deeper controversy
5:25
towards the end of 2025, when it was found to have been used for creating
5:28
sexually suggestive images of women and girls. Far from being some world-leading ChatGPT killer,
5:35
Grok was an embarrassing mess. And several xAI co-founders had already left the company.
5:40
But Musk couldn’t admit defeat. He had to keep on putting on a brave face,
5:45
acting as though this was all part of his master plan, to keep funds flowing into the project.
5:50
In early 2026, Tesla poured $2 billion of the company’s cash reserves directly into xAI.
5:56
Barely two months later, the mask finally slipped. On March 12, 2026, Musk wrote a tweet, at last
6:03
acknowledging the real state of his AI enterprise: “xAI was not built right first time around,
6:09
so is being rebuilt from the foundations up.” Almost 3 years after the company had been founded,
6:15
the man behind this grand project admitted that the whole thing was broken. It needed to
6:19
be completely undone,.just weeks after valuing the company at $250 billion to combine it with SpaceX.
6:26
For all of his posturing and proclamations, Musk had seen the data.
6:29
He knew full well that xAI’s models were miles behind those of OpenAI and other firms.
6:34
He watched as his hyped AI projects like “Grokipedia” crashed at early hurdles or were
6:40
quietly shelved. Meanwhile more and more of xAI’s top minds walked out to chase other opportunities.
6:46
Musk’s empire was starting to crack from within. His co-founders had seen it too.
6:50
They’d seen their billionaire boss talking about earth-shattering breakthroughs
6:54
and next-level technologies that would usher civilization into a whole new era.
6:58
Now, he was admitting the whole thing was broken and had to start from square one.
7:03
It was a radioactive confession with instantaneous effects.
7:07
Only a day after Musk’s tweet, two of his remaining co-founders resigned from
7:11
the company. They walked away and all the knowledge and experience
7:14
that had kept xAI running went with them. The xAI ship was taking on water and much
7:19
of the crew had already left on the lifeboats. xAI might be failing, but at least this channel
7:25
never crashes, so make sure you like, share, and subscribe. You get a front-row seat to the
7:30
chaos… without Grok insulting you. Chapter 3: The Talent Bloodbath
7:34
When he founded xAI in March 2023, Elon Musk did so alongside 11 of the
7:39
brightest minds in Silicon Valley. But as the months and years passed,
7:43
those expert architects and engineers decided that better opportunities lay elsewhere.
7:48
Kyle Kosic was the first to go, departing xAI in 2024.
7:52
He originally joined Musk from OpenAI, the biggest name in the AI industry. Musk seemingly tempted
7:58
Kosic into being part of something that he promised could be even better than OpenAI.
8:03
But a year with the company was enough for Kosic to do a U-turn, fleeing xAI and
8:08
returning to his former employer. 2025 saw two more departures.
8:12
In February, Christian Szegedy, a Google veteran who brought immense
8:15
technical experience and expertise to the xAI team, made the call to leave the company.
8:20
Next to leave was Igor Babuschkin, who originally joined xAI from Google DeepMind and served as
8:26
the company’s Chief Engineer. He announced that he was leaving the company in August,
8:30
deciding to launch his own venture firm. That was just the beginning.
8:34
The resignations ramped up in 2026. Mathematics genius and Microsoft
8:39
alumnus Greg Yang revealed in January that he was stepping down for health reasons.
8:43
The next month, Yuhuai “Tony” Wu followed suit, posting online to let followers know
8:48
he was leaving to pursue his next chapter. Wu had been one of the most significant
8:52
figures in the xAI leadership structure. A day later, xAI was dealt another blow,
8:57
with Jimmy Ba also resigning from the company. Ba is a massive name in the AI sector. He
9:02
co-authored the 2014 Adam optimization paper, the second most-cited AI paper
9:07
of all time, with over 100,000 citations. While many of the xAI co-founder departures
9:13
up to then had reportedly been quite amicable, Ba’s appeared to be different. Reports stated
9:18
that he left the company over rising tensions about demands to improve Grok’s performance.
9:23
The end of February brought about another exit, with Toby Pohlen following the example
9:28
of his fellow former co-founders. He posted a tweet in which he effectively
9:32
thanked Elon Musk for the opportunity but revealed he was leaving the company.
9:36
All of a sudden, only a fraction of the original founding team remained.
9:40
It was about to get worse. In early March, Zihang Dai
9:43
and Guodong Zhang, who formerly worked at Google Brain and DeepMind, left xAI.
9:48
Dai had been a key member of the technical team. Zhang was said to report directly to Musk and
9:53
had been put in charge of two of the company’s biggest projects: Grok Code and Grok Imagine.
9:58
By mid-March, only Manuel Kroiss and Ross Nordeen remained.
10:02
But it was only a matter of time. The reasons for their departures
10:05
have not all been shared or clearly stated, so we can only speculate as to why each of
10:09
these men chose to leave the company. Some may have had better offers from
10:13
elsewhere. Some may have grown tired of Musk’s style of leadership and failure to deliver on
10:17
his promises. Some may have simply felt that they had nothing left to give to the company, or that
10:21
their talents would be better served elsewhere. It’s not uncommon for talented tech professionals
10:25
to move about from one firm to another over time. But it is rare to see almost a dozen
10:31
different co-founders leave a company they created only a few years into its existence,
10:36
when it still, allegedly, has so much to achieve. And while there are plenty of bright minds around
10:42
Silicon Valley, people like the xAI co-founders are not easy to replace.
10:46
Every time a key figure left, xAI’s operations took a hit. Musk brought in replacements, but when
10:51
a top coder or project lead walks out, you can’t just plug someone in and pretend nothing happened.
10:58
They take talent, creativity, and experience with them, resources you can’t buy. More importantly,
11:03
they are critical in the race to dominate AI. And while Musk was busy tweeting,
11:08
making more promises he couldn’t keep, and watching the resignations pile up,
11:11
his rivals weren’t slowing down. Chapter 4: The War Musk Can’t Win
11:16
While Elon Musk’s allies were abandoning him, one of the Tesla boss’s biggest rivals was
11:21
making moves. Mark Zuckerberg was offering jaw-dropping $300 million contracts over
11:27
4 years to lure the best and brightest minds in the AI industry to join his team at Meta.
11:32
At OpenAI and Google, too, Sam Altman and Jeff Dean have been doing everything in
11:37
their power to hunt down the engineers, designers, mathematicians, and computer
11:41
scientists who can lead their projects to glory. The people joining these firms aren’t “selling
11:45
out” or simply looking for the biggest paydays. They’re seeking stability.
11:50
They want to be part of long-term projects with clear goals, vision,
11:53
and structure, operating under project heads. They want bosses who know how to
11:57
strike the balance between ambition and realism. Musk may have his world-leading supercomputer.
12:04
He may still be able to boast about the number of GPUs powering Grok or the amount
12:08
of megawatts his AI data centers consume. But all of that means little without
12:13
an equally world-leading team of talented people to run the show.
12:17
Meanwhile, over at Google, OpenAI, Meta, and elsewhere, multifaceted super-teams of
12:22
AI engineers and innovators are being assembled, all while infrastructure is steadily improving.
12:27
Across the US - and even in China - AI rivals are catching up. They’re building
12:31
massive data centers, scaling their teams, and improving their technology in smart,
12:35
strategic ways. Meanwhile, Musk remains stuck in his Grok Ghost Town, watching the competition
12:41
show him exactly how it should be done. And with every passing day, the odds are
12:45
stacked increasingly in favor of those firms. xAI’s chances of success are slipping away.
12:50
Because companies like OpenAI and Google don’t play around when it comes to talent.
12:55
They have huge teams, thousands of PhDs, billions of dollars invested in research
13:00
and development. They bring years of combined experience, unparalleled levels of expertise,
13:05
all working together towards clear, common goals. These companies have faith in the process.
13:11
They know that greatness can’t be rushed, and that a little time and patience are sometimes
13:16
needed to bring about the biggest, most important breakthroughs. They build their
13:20
AI models block by block, iterating gradually but convincingly, building slowly but surely
13:25
towards something better than before. Musk’s “skeleton crew” staff and “move
13:29
fast and break things” attitude were never going to be enough to compete with that.
13:33
He thought he could sprint to the top, using his vast fortune to brute force his way to success.
13:38
And he thought his team of co-founders would remain loyal to the cause, every step of the way.
13:43
But that’s not what happened. It took some a little longer than others,
13:48
but one by one, xAI’s creators grew tired and disillusioned with the project. They
13:53
lost faith in Musk, lost belief in Grok, and lost patience with the entire company.
13:58
Knowing that every year of their careers is precious, they each made the call to
14:01
pursue other options, undermining Musk’s boasts about xAI being the best name in the industry.
14:07
The Tesla boss’s erratic behavior certainly hasn’t helped things.
14:10
Musk is famous for hyper-aggressive timelines, hardcore work patterns, near-impossible
14:15
objectives, and constant moving of the goalposts. And that kind of approach might have worked out
14:20
for him at Tesla, and even, to some extent, at X. But AI research, by its very nature, requires time
14:26
and patience. It’s not something Musk can speed up or solve by snapping his fingers or flinging
14:31
more money at the problem. And when you’re working with people at the top of their game,
14:35
with incredible intelligence and amazing ideas floating around their minds, they have to be given
14:40
the freedom to put their skills to best use. It seems, at least from the outside,
14:44
that Musk failed to do that. He is the king of an empty castle, the captain
14:49
of a sinking ship, the face of a failing brand. All his closest subjects have left him.
14:54
While a few new faces have joined xAI, it may be only a matter of time before they,
14:58
too, follow their predecessors and opt out of the company. Even after the megamerger with SpaceX,
15:04
xAI still lags majorly behind its rivals when it comes to real human brainpower.
15:09
It may have the funding to survive for many more years and the tech to remain
15:14
semi-relevant in the sector for the foreseeable future, but it would take a Herculean effort to
15:19
turn things around and even come close to catching up with its main rivals.
15:23
Instead, what is likely to happen is that OpenAI, Meta, Anthropic, and other firms will surge ahead,
15:29
leaving Musk even deeper in the dust. Musk may have the money, the tech,
15:33
and the ambition, but is it enough to turn things around or is he already falling too far behind? To
15:38
see the full story watch “The Ugly Truth About Elon Musk.” Or click on this video instead.
