1
00:00:05,070 --> 00:00:08,630
Cerebres upsizes IPO to $4.8 billion.

2
00:00:08,880 --> 00:00:12,039
$4.8 billion now earmarked for wafer pre-purchases

3
00:00:12,039 --> 00:00:13,939
before the first lockup even starts.

4
00:00:14,319 --> 00:00:19,700
That volume forces TSMC to reallocate capacity away from smaller AI chip designers

5
00:00:19,879 --> 00:00:21,360
who cannot match the commitment.

6
00:00:21,760 --> 00:00:25,239
Nvidia will have to decide whether to defend margins or flood the market

7
00:00:25,239 --> 00:00:29,399
with discounted Blackwell units just to keep key accounts from switching over.

8
00:00:29,760 --> 00:00:32,959
Smaller players face a choice between expensive spot buys or

9
00:00:32,959 --> 00:00:37,099
delayed tapeouts that push their product launches well into 2027.

10
00:00:38,790 --> 00:00:42,049
OpenAI launches $4 billion dollars deployment company.

11
00:00:42,349 --> 00:00:46,270
Embedding full-time AI engineers inside client environments changes the

12
00:00:46,270 --> 00:00:50,409
revenue mix from pure API calls to recurring services fees

13
00:00:50,509 --> 00:00:52,409
that carry lower gross margins.

14
00:00:52,750 --> 00:00:56,930
That shift immediately pressures Anthropic to stand up its own deployment team

15
00:00:57,049 --> 00:01:00,069
or watch mid-market accounts migrate to the new offering.

16
00:01:00,409 --> 00:01:03,029
The acquisition of Tomorrow accelerates the timeline,

17
00:01:03,369 --> 00:01:06,169
but it also exposes OpenAI to fixed labor

18
00:01:06,169 --> 00:01:09,750
costs that pure token-based margins never carried before.

19
00:01:10,050 --> 00:01:12,529
Expect enterprise contracts to include headcount

20
00:01:12,529 --> 00:01:14,849
minimums within the next two quarters.

21
00:01:16,860 --> 00:01:21,440
Anthropic ARR hits $44 billion after 80 times growth

22
00:01:21,620 --> 00:01:22,519
in a single year,

23
00:01:22,540 --> 00:01:25,000
usually signals either a massive new market

24
00:01:25,150 --> 00:01:28,610
or unsustainable discounting just to land the biggest logos.

25
00:01:29,010 --> 00:01:30,889
Dario's number points to the latter.

26
00:01:31,180 --> 00:01:36,080
OpenAI now faces pressure to launch its own deployment services at similar scale

27
00:01:36,260 --> 00:01:39,860
or lose the same enterprise pipeline that Anthropic just captured.

28
00:01:40,300 --> 00:01:43,279
The real test arrives when those contracts hit renewal

29
00:01:43,430 --> 00:01:47,620
and the burn rate for Frontier model training collides with services overhead

30
00:01:47,860 --> 00:01:49,779
that investors have not modeled yet.

31
00:01:51,379 --> 00:01:54,489
Pentagon signs AI deals excluding Anthropic.

32
00:01:54,860 --> 00:01:57,010
The supply chain explanation masks a

33
00:01:57,010 --> 00:02:00,120
deeper mismatch between Anthropic's public safety charter

34
00:02:00,300 --> 00:02:03,620
and Pentagon requirements for direct weapons system integration.

35
00:02:04,110 --> 00:02:08,389
OpenAI and Google now gain exclusive access to classified data streams

36
00:02:08,580 --> 00:02:12,160
that will shape their next model fine tunes ahead of everyone else.

37
00:02:12,500 --> 00:02:16,559
Anthropic must either carve out a defense subsidiary with separate governance

38
00:02:16,779 --> 00:02:21,039
or accept permanent exclusion from the largest single AI buyer in the world.

39
00:02:21,380 --> 00:02:23,929
Either path changes their roadmap and valuation

40
00:02:23,929 --> 00:02:26,279
trajectory within the next 18 months.

41
00:02:27,720 --> 00:02:30,710
Google finds first AI built zero day exploit.

42
00:02:31,039 --> 00:02:34,320
Security teams now face exploits that AI wrote from scratch,

43
00:02:34,440 --> 00:02:35,619
not just polished.

44
00:02:35,839 --> 00:02:39,940
The real shift hits when every mid-tier group starts generating these at scale.

45
00:02:40,160 --> 00:02:43,419
Expect zero day volume to double inside 12 months.

46
00:02:43,660 --> 00:02:46,300
That compresses the patch window from weeks to days

47
00:02:46,440 --> 00:02:51,320
and forces SOC tools to add real-time model scanning or lose their edge entirely.

48
00:02:51,740 --> 00:02:54,360
One admin tool breach already shows the pattern,

49
00:02:54,589 --> 00:02:58,639
and the next wave will target multi-factor flows across the board without warning.

50
00:02:59,630 --> 00:03:02,100
Claude code adds agent view dashboard.

51
00:03:02,419 --> 00:03:05,880
Parallel agent sessions just became manageable from one screen,

52
00:03:06,100 --> 00:03:10,380
but the real test comes when those agents edit overlapping files at the same time.

53
00:03:10,740 --> 00:03:12,690
Merge conflict rates will spike unless

54
00:03:12,690 --> 00:03:15,330
the dashboard adds automatic coordination hooks.

55
00:03:15,660 --> 00:03:16,369
Without that,

56
00:03:16,660 --> 00:03:19,449
parallel work turns into coordination debt fast.

57
00:03:19,779 --> 00:03:23,770
Rivals shipping single agent tools now have 6 months to match the orchestration

58
00:03:23,770 --> 00:03:26,779
layer before .NET teams standardize on Claude's

59
00:03:26,779 --> 00:03:29,259
view and stop switching contexts altogether.

60
00:03:30,110 --> 00:03:31,389
Workday CEO

61
00:03:31,630 --> 00:03:32,880
AI replaces labor,

62
00:03:32,990 --> 00:03:34,029
not software.

63
00:03:34,270 --> 00:03:37,490
Enterprise software budgets rarely shrink when AI arrives,

64
00:03:37,669 --> 00:03:39,610
yet this stance keeps the core platform

65
00:03:39,610 --> 00:03:42,169
untouched while trimming the people layer instead.

66
00:03:42,589 --> 00:03:44,330
Large workday customers therefore

67
00:03:44,330 --> 00:03:47,070
redirect transformation dollars toward internal

68
00:03:47,070 --> 00:03:49,990
process bots rather than rip and replace projects.

69
00:03:50,380 --> 00:03:54,070
That squeezes the services arms at Salesforce and Oracle first,

70
00:03:54,309 --> 00:03:56,830
pushing them to cut implementation headcount before

71
00:03:56,830 --> 00:03:58,889
their license streams feel any pressure.

72
00:03:59,369 --> 00:04:03,089
The pattern repeats across HCM suites as buyers realize

73
00:04:03,089 --> 00:04:05,929
the existing stack already contains the hooks they need.

74
00:04:07,899 --> 00:04:10,779
Re-sharper adds AI agents to Visual Studio.

75
00:04:11,020 --> 00:04:13,660
.net developers gain agents that refactor

76
00:04:13,660 --> 00:04:16,320
across solutions without leaving Visual Studio.

77
00:04:16,540 --> 00:04:18,980
Yet the second order effect appears in how quickly

78
00:04:18,980 --> 00:04:21,420
jet brains can lock in that workflow data.

79
00:04:21,898 --> 00:04:25,178
Microsoft now has until the full 2026.2

80
00:04:25,178 --> 00:04:28,039
release to match agent depth inside co-pilot

81
00:04:28,158 --> 00:04:32,209
or risk losing daily driver status among enterprise.net shops.

82
00:04:32,658 --> 00:04:36,158
Teams testing the early access already report fewer handoffs,

83
00:04:36,299 --> 00:04:39,329
accelerating the shift away from piecemeal extensions.

84
00:04:39,658 --> 00:04:41,699
The lock-in potential grows each time an

85
00:04:41,699 --> 00:04:44,079
agent learns the team's specific patterns.

86
00:04:45,799 --> 00:04:48,480
AI product searches surge across ages.

87
00:04:48,839 --> 00:04:53,100
Which age group actually converts fastest once AI surfaces the links?

88
00:04:53,440 --> 00:04:55,459
Younger users click through immediately,

89
00:04:55,600 --> 00:04:58,119
while older demographics still pause to verify

90
00:04:58,119 --> 00:05:00,440
seller ratings and read reviews manually,

91
00:05:00,760 --> 00:05:03,040
creating a measurable lag that retailers must now.

92
00:05:03,160 --> 00:05:05,750
Engineer around with separate trust signals.

93
00:05:06,160 --> 00:05:08,519
This split will push mid-market platforms to

94
00:05:08,519 --> 00:05:12,859
deploy age-specific agent interfaces by early 2026

95
00:05:13,000 --> 00:05:15,739
or lose share to Amazon's unified experience.

96
00:05:16,079 --> 00:05:20,440
Older users will demand those extra steps or abandon the flow entirely.

97
00:05:21,089 --> 00:05:23,760
OpenAI shares cyber model with EU.

98
00:05:24,049 --> 00:05:26,089
Brussels now holds leverage that no other

99
00:05:26,089 --> 00:05:28,869
jurisdiction has secured from a frontier lab.

100
00:05:29,130 --> 00:05:33,570
OpenAI's move gives regulators direct testing rights on live cyber tools,

101
00:05:33,850 --> 00:05:37,709
something competitors have avoided by staying silent on regional access.

102
00:05:38,005 --> 00:05:41,744
This forces Anthropic to either open mythos under similar terms

103
00:05:41,924 --> 00:05:44,144
or accept a shrinking European footprint

104
00:05:44,285 --> 00:05:47,005
as enterprise customers migrate toward compliant

105
00:05:47,005 --> 00:05:48,704
options within the next 12 months.

106
00:05:49,165 --> 00:05:50,704
Labs that delay compliance

107
00:05:50,845 --> 00:05:54,825
will hand OpenAI an uncontested lead in regulated sectors.

108
00:05:56,559 --> 00:05:59,579
AWS Bedrock Agent Corp previews payments.

109
00:05:59,920 --> 00:06:02,600
Agents handling payments will process their first million

110
00:06:02,600 --> 00:06:06,470
transactions before most enterprises finish their security reviews.

111
00:06:06,880 --> 00:06:10,940
Bedrock's preview lets agents bypass human approval loops entirely

112
00:06:11,079 --> 00:06:13,790
when calling paid APIs through Stripe and Coinbase.

113
00:06:14,369 --> 00:06:16,459
Watch what happens to the API margin

114
00:06:16,570 --> 00:06:19,570
when this hits enterprise SLAs and forces

115
00:06:19,570 --> 00:06:22,399
providers to pre-authorize autonomous spend limits,

116
00:06:22,410 --> 00:06:25,470
or lose agent traffic to cheaper open endpoints.

117
00:06:25,730 --> 00:06:28,309
The providers that set clean spend limits first

118
00:06:28,420 --> 00:06:31,209
will capture the initial wave of agent-driven traffic.

119
00:06:32,720 --> 00:06:35,239
AI pushes data centers to Europe's edge.

120
00:06:35,609 --> 00:06:36,769
Power availability,

121
00:06:36,970 --> 00:06:38,160
not land or fiber,

122
00:06:38,410 --> 00:06:42,769
now dictates where the next wave of AI data centers can actually open in Europe.

123
00:06:43,059 --> 00:06:47,260
The projected $7 trillion in spend will concentrate first around

124
00:06:47,260 --> 00:06:51,519
existing hydro and nuclear sites rather than new greenfield builds.

125
00:06:51,850 --> 00:06:56,320
That concentration accelerates renewable interconnection cues in Scandinavia,

126
00:06:56,500 --> 00:06:58,600
while leaving Southern Europe grids strained

127
00:06:58,739 --> 00:07:04,160
and pushing some inference workloads back to centralized US regions by late 2026.

128
00:07:04,459 --> 00:07:08,899
Operators ignoring power constraints will hit permitting walls by 2027.