DATE: 6/14/83
SUBJECT: Nintendo
As we discussed with Dave Stubben yesterday, I am to remove
myself from any further involvement with the Nintendo project
..... and Dave has indicated that you are the one to take it
over.
I have provided you with a copy of my file containing my previous
memos on the subject (with corrections); the approach letter from
Nintendo; all of the schematics and mechanical drawings of the
Nintendo machine which we have received to date; and Ed Levy's
mechanical drawings which attempt to fit the Nintendo machine
into the 2100 plastic.
We spent the latter part of the afternoon yesterday discussing
the history of this deal and what needs to be done next by both
sides. Let me review those points here and expand upon them for
your reference.
Mr. Henricks received a letter from Mr. Arakawa and Mr. Lincoln
of Nintendo America on April 4, 1983. In that letter, Nintendo
provided us with some preliminary specifications on their new
home video game machine. A couple of days later they came to meet
with Mr. Kassar to explore whether Atari had any interest in this
product. In addition to Messers Arakawa, Lincoln, Henricks, and
Kassar; Messers Groth, Malloy, Moone, Bruehl, Ruckert, and myself
were invited to attend. Mr. Malloy and I framed some of the
initial questions which needed to be answered about the
capabilities of the machine; and Mr. Lincoln promised to get the
answers to me within a few days. Those answers were sufficiently
intriquing to Mr. Groth that Alan Henricks, Dave Remson, and I
were asked to travel to Kyoto immediately to see their TTL
emulator in action and get more details about the final product.
On the 11th of April, 1983, we met with Nintendo at their
headquarters in Kyoto. By happenstance (fortunate or
unfortunate), a large contingent of Atari executives were in the
Far East for other reasons .... and they all decided to come to
Kyoto to have a look, too. In attendance from Atari were Messers.
Bruehl, Moone, Malloy, Lynch, Hennick, Mitoh, Henricks, Remson,
and myself. Attending on Nintendo's behalf were Messers. Yamauchi
(President), Takeda (Manager of R&D, Coin-Op), Arakawa (Pres.
Nintendo America), Lincoln (internal attorney for Nintendo
America), Uemura (Manager of R&D, Consumer Products), Todori
(Export Manager), and two of their electrical engineers.
We were shown working (but not complete) versions of Donkey Kong
Junior and Popeye running with only minor display glitches on
their TTL emulator. A VHS video tape (without sound) of that demo
is attached to this memo. Please keep in mind that the actual TV
image is significantly better than could be captured on tape. In
fact, there is a noticable difference when viewing the composite
video output on a monitor as opposed to the RF output on a
standard TV receiver.
At that time, Nintendo had only just received their 1st pass
silicon (with some bugs) and were not able to show us a fully
assembled and working prototype. My memo of 4/16/83 (with
corrections) describes what we saw and were told in that meeting.
On 4/15/83, Messers Kassar, Groth, Moone, Bruehl, Paul, Henricks,
Remson, and myself met in Mr. Groth's office to view the videotape
and discuss what we had learned from the meeting on the 11th and
what we knew to-date on the MARIA chip being developed by General
Computer. As both systems were seen as being in the same price
range with graphics capabilities superior to the 2600 and
comparable (and in some features, superior) to the 5200, it was
felt that we needed to see what could be done with both machines
for an intermediate priced game machine .... the 3600.
I was asked to become as completely informed about the MARIA chip
as possible so that a reasoned choice could be made between the
two machines. To that end, I have spoken with the folks at General
Computer several times by telephone and have made two trips to
their offices in Cambridge, Mass. It appears to be a superior
machine, but the MARIA chip is not yet finished. First silicon is
not expected until mid-July (if there are no further schedule
delays). Also, since this chip is a VTI device there is some
question as to the manufacturability/testability/cost of the chip.
In other words, it will not be until mid-July (mid-August if the
first silicon is faulty) that we will be able to make a fully
informed choice between the Nintendo and the MARIA machines.
Therefore, it was decided by Executive Management that in any
negotiations with Nintendo we would need to string out the
signing until at least mid-July.
We were committed to respond quickly to Nintendo, however, as to
whether we were interested or not. So, Alan Henricks did contact
Nintendo with the word that we were interesting in continuing the
discussions; and the next negotiating meeting was arranged for May
17th in Kyoto. Skip Paul and Alan Henricks were to represent
Atari. Two or three days before that meeting, Nintendo informed
Mr. Henricks that they would be having their senior engineering
managers present in the negotiations, and Nintendo requested my
attendance as well.
That negotiating session began with a statement from Mr. Yamauchi
as to the terms and conditions which he demanded, namely:
1. that Atari would purchase the assembled and tested
main pc board for the FCS from Nintendo, for sale
outside of Japan. Nintendo would sell the FCS on its
own in Japan.
2. after some minimum purchase of assembled and tested pc
boards, we would be able to buy the 2 custom chips
from Nintendo without having to have Nintendo assemble
them into the final unit.
3. that Nintendo would only disclose the electrical specs
for the PPU and CPU, the circuit diagram of the FCS
system, the test programs, and the "cassette" specs
(meaning the ROM cartridge and cartridge edge
connector specs).
4. that there would be no disclosure to Atari of the
programming specs for the PPU and the CPU.
5. that Nintendo would program titles of our choice for
the FCS system and would sell us the assembled and
tested, unlabeled ROM carts at 1,500 Yen each FOB
Japan for retail sale by Atari. The minimum quantity
required by Nintendo per title would be 100,000 units
and at that level there would be no fee for
non-recurring engineering/programming expenses.
6. that Atari would hereby obtain a "right of 1st
refusal" on future Nintendo coin-op titles for use
worldwide (outside of Japan) only for the Nintendo FCS
system .... again, by programming and manufacturing
those carts themselves for sale to us.
7. that the cost of the assembled and tested main pc
board would be higher than the 5,300 Yen quoted
earlier to cover the cost of FCC compatibility. Also,
that the resulting new pc board would not fit into the
plastic being used by Nintendo for this unit in Japan.
By the time we finished the negotiations on that trip (5/17 -
5/20), the deal was changed to be as follows:
A. Nintendo would disclose all items called for in my
memo of 5/13/83 (to Henricks and Paul) except for item
13., namely, the LSI tapes for chip fabrication. This
disclosure would take place upon signing of the deal.
All items which are originally in Japanese are to be
furnished to us both in Japanese and in English.
B. Upon signing the deal, Nintendo would reassure Atari
about the source of supply of the 2 custom chips.
C. Any increases in the cost of the main pc board due to
FCC compliance will be a straight cost pass through
(no additional profit to Nintendo).
D. Atari and Nintendo would work together to attempt to
legal protect the CPU and PPU designs.
E. Nintendo would receive $5. Mil upon signing as an
advance against future payments.
F. Atari would have to commit to a minimum purchase of 2
million hardware units (some mixture of assembled and
tested pc boards and CPU/PPU chip sets) over the term
of the contract.
G. The term of the contract would be 4 years with a 4
year option to renew.
H. Nintendo would receive an additional $3.5 Mil in a
line of credit as an advance upon future payments upon
delivery of the 1st production-ready prototype of the
PAL West Germany version of the FCS (no later than
1/1/84). Similarly, an additional advance of $1.5 Mil
for SECAM.
I. The 2 million unit commitment would be broken-up into
1 million NTSC, 700,000 PAL, and 300,000 SECAM. If
Atari goes over in one catagory, it would directly
reduce our requirement in any other catagory of our
choice. As Skip Paul likes to put it, "cross-
collateralization is the key!"
J. Nintendo would commit to produce 100,000 units of the
assembled and tested pc board by August 31st if the
new pc design (to include FCC and to fit whatever
plastic we choose) can be completed by Nintendo
and approved by Atari by July 20th. In essence,
unlimited quanities (in excess of 1 million/month)
thereafter .... upon 3 months notice from Atari.
K. Atari will have the right to program for this system
with the full assistance of Nintendo.
L. Nintendo will, in the interests of expediency for this
Christmas season, program 4 Atari titles of our
choice. Source and object code which meets our
satisfaction (with respect to basic design, tuning,
and bug-free) to be delivered to us no later than
Sept. 1, 1983. The fee would be $100,000./title or
no non-recurring engineering fees would be charged as
long as we buy a minimum of 100,000 carts.
M. Carts would cost us 1,500 Yen/cart if in plastic but
unlabeled or 1,350 Yen if not in plastic (F.O.B.
Japan). Rate of production would be max. 5,000
units/week/title
